Google has launched the full version of Gemini 2.5 Flash, its advanced AI image editing model, expanding its availability to developers and users worldwide. The rollout marks a major update to Google’s AI image ecosystem, bringing enhanced editing capabilities and new creative tools to the public. Among the most notable additions are expanded aspect ratio options and an image-only output mode in its image generator app, Nano Banana. The update enables users to produce visuals in specific shapes tailored for social media, presentations, or cinematic frames, offering more control over creative outcomes. Gemini 2.5 Flash now supports a wide range of aspect ratios—landscape (21:9, 16:9, 4:3, 3:2), square (1:1), portrait (9:16, 3:4, 2:3), and flexible formats (5:4, 4:5)—making it suitable for a diverse set of use cases from digital marketing to entertainment design.
The new image-only output option allows users to generate visuals without accompanying text responses. Previously, Gemini’s default mode provided both text and image outputs together, which often cluttered workflows for those seeking purely visual results. Now, by adjusting response_modalities to [‘Image’], creators and developers can streamline outputs to focus solely on generated content. Gemini 2.5 Flash first appeared in late August as a preview through the Gemini API and Google AI Studio, where developers began experimenting with it to build custom AI-powered applications. The technology’s flexibility allows users to design their own image editing tools or utilize demo apps created by Google engineers. These include Past Forward, an app that places a user’s portrait into different historical eras, and Fit Check, which can alter clothing styles and poses.
In addition to editing still images, Gemini 2.5 Flash interprets natural language commands, enabling intuitive photo manipulation through simple prompts. Within Google’s Pixshop app, for instance, users can instruct the model to remove people from an image, blur a background, or colorize old photos. Gemini Co-Drawing combines drawing input with AI refinements—users can sketch a concept and then direct the model to enhance or complete it with text instructions. The model also incorporates real-world understanding, allowing it to recognize and label objects, measure proportions, or insert related items in images. In the Home Canvas app, users can merge multiple photos seamlessly by dragging an object image into a scene, with the model automatically blending the new element naturally into its environment.
Gemini 2.5 Flash is now available for testing in Google AI Studio, where developers can explore pre-built apps or build their own using a simple text-based interface. Through the Gemini API, the model can also be embedded into enterprise systems via Google Cloud’s Vertex AI platform. While image generation through AI Studio remains free, high-volume or production users can switch to pay-as-you-go pricing, with costs set at $0.039 per image and $30 per million tokens for multimodal outputs. Beyond Google’s ecosystem, OpenRouter has introduced Gemini 2.5 Flash as its first image model, while fal.ai and Adobe have integrated it into their developer and creative tools. This expansion highlights how AI image technology is becoming more accessible and adaptable, giving users greater freedom to integrate generative capabilities into everyday creative and professional workflows.
Follow the SPIN IDG WhatsApp Channel for updates across the Smart Pakistan Insights Network covering all of Pakistan’s technology ecosystem.
